Public Release: 

UT Arlington research looks to unlock connections among social network data

Big data

University of Texas at Arlington


IMAGE: This is Gautam Das, professor of the Computer Science and Engineering Department in the UT Arlington College of Engineering. view more

Credit: UT Arlington

A UT Arlington computer science and engineering professor has won a $450,000 Army Research Office grant to develop efficient analytic techniques for combining and understanding the data stored in online social networks.

Gautam Das, professor of Computer Science and Engineering in the College of Engineering, said the grant is to find "implicit edges" in the uses of social networks. Implicit edges connect two seemingly unrelated occurrences on a social media platform. In contrast, explicit edges represent the more obvious relationships on the social media platform, such as friends, followers and contacts.

"We are to determine which are the more promising directions to pursue when seeking more hidden data," said Das, who is director of UT Arlington's Database Exploration Laboratory or DBXLab. "We'll look at who is acquainted or are followers of one person or another across social networks."

The research could be used to improve military intelligence, make electronic commerce more efficient, help determine pharmaceutical side effects of certain drugs and help identify topics generating media buzz and public interest, among other applications.

Khosrow Behbehani, dean of the College of Engineering, said Das's work expands UT Arlington's increasing expertise in the area of "Big Data."

"We all are generating increasing amount of data through our social interactions on the internet, " Behbehani said. "Each social network user may have significant volumes of such data. Uncovering the interdependencies and relationships between the data can be valuable in a number of areas."

Each node or edge may contain a significant amount of data in the form of Facebook posts, tweets or retweets. Das said a third-party analyst, like an Army intelligence agent, might see the link within that data.

"Seeing the relationships is difficult to do because third-party users, like us, don't have full access to the entire database," Das said. "Our access would be limited by the web interface."

Das said an important feature of the hidden-access model the team will create is that it enables the researchers to construct, model and understand "virtual" social networks that are built upon not one, but multiple, web data sources.

For example, from the view of an Army intelligence agency, a real-world social network formed by a group of suspected terrorists can be considered a virtual social network. This network does not reside on any single online social network website like Facebook or Twitter, but instead leaves traces in a variety of web data sources, like forums, exchanges on online social networks and even comments on certain stories on the web.

Since the total implication of such a virtual network is unknown to anyone - except perhaps a few terrorists in the network - existing techniques cannot be applied, Das said.

The proposed techniques could leverage the types of information revealed through different web sources to gain insights into such a virtual social network.

Das has long had research interests in big data exploration. Before joining UT Arlington a decade ago, Das worked at Microsoft Research, Compaq Corp. and the University of Memphis. This project will be in collaboration with Professor Nan Zhang of George Washington University, a long-time collaborator of Das and a former faculty member of UT Arlington's CSE department.


About UT Arlington

The University of Texas at Arlington is a comprehensive research institution of more than 40,000 students worldwide and the second largest institution in The University of Texas System. The Chronicle of Higher Education ranked UT Arlington as the seventh fastest-growing public research university in 2013. U.S. News & World Report ranks UT Arlington fifth in the nation for undergraduate diversity. Visit to learn more. Follow #UTAdna on Twitter.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.